How do I encode arbitrary data to XML using Java 1.4 and SAX?

Question

We use SAX to parse XML because it does not require the entire XML document to be read into memory in order to parse a single value. I read many articles that insisted SAX can only be used to parse/decode XML and not create it. Is this true?

Answer 1

No, it isn't true, you can encode XML to any Writer in Java using something similar to:

char[] ch;
AttributesImpl atts = new AttributesImpl();
Writer writer = new StringWriter();
StreamResult streamResult = new StreamResult(writer);
SAXTransformerFactory tf = (SAXTransformerFactory) TransformerFactory.newInstance();

// SAX2.0 ContentHandler
TransformerHandler transformerHandler = tf.newTransformerHandler();

Transformer serializer = transformerHandler.getTransformer();

serializer.setOutputProperty(OutputKeys.ENCODING, "UTF-8");
// serializer.setOutputProperty(OutputKeys.DOCTYPE_SYSTEM, "nodes.dtd");
serializer.setOutputProperty(OutputKeys.INDENT, "yes");

transformerHandler.setResult(streamResult);

transformerHandler.startDocument();

atts.clear();
// atts.addAttribute("", "", "xmlns", "CDATA", "http://www.example.com/nodes");
// atts.addAttribute("", "", "xmlns:xsi", "CDATA", "http://www.w3.org/2001/XMLSchema-instance");
// atts.addAttribute("", "", "xsi:schemaLocation", "CDATA", "/nodes.xsd");
transformerHandler.startElement("", "", "node_list", atts);

// displayName element
if (displayName != null) {
 transformerHandler.startElement("", "", "display_name", null);
 ch = displayName.toCharArray();
 transformerHandler.characters(ch, 0, ch.length);
 transformerHandler.endElement("", "", "display_name");
}

// nodes element
transformerHandler.startElement("", "", "nodes", null);

atts.clear();
atts.addAttribute("", "", "node_type", "CDATA", "sometype");
transformerHandler.startElement("", "", "node", atts);

ch = node.getValue().toCharArray();
transformerHandler.startElement("", "", "value", null);
transformerHandler.characters(ch, 0, ch.length);
transformerHandler.endElement("", "", "value");

transformerHandler.endElement("", "", "node");

transformerHandler.endElement("", "", "nodes");

transformerHandler.endElement("", "", "node_list");

transformerHandler.endDocument();

String xml = writer.toString();

Answer 2

The SAX handler interfaces were designed to be easy to implement. It's easy to write a class with similar (perhaps wrapping a SAX interface) to make it easy to call - chaining, remembering which element to close, easier attributes, etc.

Answer 3

see http://code.google.com/p/jlibs/wiki/XMLDocument

It supports creating xml document using SAX. both memory efficient and user friendly api